All Downloads are FREE. Search and download functionalities are using the official Maven repository.

org.elasticsearch.common.lucene.all.AllField Maven / Gradle / Ivy

There is a newer version: 7.10.2_1
Show newest version
/*
 * Licensed to Elasticsearch under one or more contributor
 * license agreements. See the NOTICE file distributed with
 * this work for additional information regarding copyright
 * ownership. Elasticsearch licenses this file to you under
 * the Apache License, Version 2.0 (the "License"); you may
 * not use this file except in compliance with the License.
 * You may obtain a copy of the License at
 *
 *    http://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing,
 * software distributed under the License is distributed on an
 * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
 * KIND, either express or implied.  See the License for the
 * specific language governing permissions and limitations
 * under the License.
 */

package org.elasticsearch.common.lucene.all;

import org.apache.lucene.analysis.Analyzer;
import org.apache.lucene.analysis.TokenStream;
import org.apache.lucene.document.Field;
import org.apache.lucene.document.FieldType;
import org.apache.lucene.index.FieldInfo.IndexOptions;
import org.elasticsearch.ElasticsearchException;

import java.io.IOException;
import java.io.Reader;

/**
 *
 */
public class AllField extends Field {

    private final AllEntries allEntries;

    private final Analyzer analyzer;

    public AllField(String name, AllEntries allEntries, Analyzer analyzer, FieldType fieldType) {
        super(name, fieldType);
        this.allEntries = allEntries;
        this.analyzer = analyzer;
    }

    @Override
    public String stringValue() {
        if (fieldType().stored()) {
            return allEntries.buildText();
        }
        return null;
    }

    @Override
    public Reader readerValue() {
        return null;
    }

    /** Returns the {@link AllEntries} containing the original text fields for the document. */
    public AllEntries getAllEntries() {
        return allEntries;
    }

    @Override
    public TokenStream tokenStream(Analyzer analyzer, TokenStream previous) throws IOException {
        try {
            allEntries.reset(); // reset the all entries, just in case it was read already
            if (allEntries.customBoost() && fieldType().indexOptions().compareTo(IndexOptions.DOCS_AND_FREQS_AND_POSITIONS) >= 0) {
                // TODO: we should be able to reuse "previous" if its instanceof AllTokenStream?
                // but we need to be careful this optimization is safe (and tested)...
                
                // AllTokenStream maps boost to 4-byte payloads, so we only need to use it any field had non-default (!= 1.0f) boost and if
                // positions are indexed:
                return AllTokenStream.allTokenStream(name, allEntries, analyzer);
            } else {
                return analyzer.tokenStream(name, allEntries);
            }
        } catch (IOException e) {
            throw new ElasticsearchException("Failed to create token stream");
        }
    }
}




© 2015 - 2024 Weber Informatics LLC | Privacy Policy